Effects of different CBCT acquisition protocols on radiation dose and observer agreement in the assessment of impacted mandibular third molars.

This study aimed to evaluate the effects of different CBCT (Cone-beam computed tomography) protocols defined by device type, field of view (FOV), and dose settings on organ-specific radiation doses and inter-/intra-observer agreement in the assessment of impacted mandibular third molars. A total of 168 CBCT scans (NewTom 7G and NewTom Go) were retrospectively analyzed. Each device included six protocols (different FOVs and dose levels: low, regular, best). Organ and effective doses were calculated using the Monte Carlo PCXMC 2.0 Rotation software. Three observers independently assessed five anatomical parameters related to surgical risk. Inter- and intra-observer agreement was analyzed using Fleiss' and Cohen's kappa. Effective doses ranged from 8.65 to 376.96 µSv, a more than 40-fold difference between the lowest- and highest-dose protocols. NewTom Go protocols with limited FOV and low/regular doses yielded the lowest effective doses without a corresponding loss of observer agreement. Observer agreement was substantial to almost perfect for most parameters (kappa: 0.541-0.915), although agreement for periodontal ligament visibility and root resorption was partly driven by shared "not visible"/"uncertain" ratings rather than confident structure identification. As no clinical or surgical reference standard was used, these findings reflect observer agreement and diagnostic interpretability rather than confirmed diagnostic accuracy. Limited FOV and optimized dose settings reduced radiation exposure while preserving observer agreement, supporting ALADAIP (As Low As Diagnostically Acceptable being Indication-oriented and Patient-specific) principles.